[TowerTalk] Tuners

[Jim Thomson]

Date: Mon, 29 Nov 2010 16:49:42 +0000
From: Steve Hunt <steve at karinya.net>
Subject: Re: [TowerTalk] Tuners

That's a useful tuner simulation; but **DO** remember it assumes a 
constant Q for the inductor. Here's how a real roller inductor Q 
measures at different inductance settings on different bands:

http://www.karinya.net/g3txq/temp/rollercoaster_q.png

If you want to get meaningful loss results with that simulator, be 
prepared to manually change the default Q value as you change bands and 
change inductance values.

73,
Steve G3TXQ

##  Per this chart,  from 20-10m, the results are abysmal.   The roller in question
is a 18 uh unit, with a tapered pitch at one end. ..and uses 12 ga wire.

##  a couple of point's here.   Using Dr Bing's  RF  software  for the ideal T match
values for the low bands,  say 160m,  with a 50 ohm input, and a 10 ohm resistive output,
[zero reatance], it spits out very high values for the 2 x caps,  and a very low value for the coil. 
6289pf  for C1 [109 volts peak, and .4 w diss]...and   4171 pf for C2.
[359 v peak and 2.6 W]  Coil is 1.98 uh.   With a coil Q = 125, the
coil dissipates 28 watts.   I used a Q= 1200  for both caps. [1500w CXR]

##  Now if we change the phase shift, and alter the value of the 3 x components,
we end up with C1= 431pf. C2= 941pf  Coil = 5.71uh.   Coil diss increases to 
a whopping 143 watts. 

##  Even Palstar  says their tuners are... "optimized for the lower bands"   Presumably
they mean the typ 500 pf cap has a high min C, and / or the  coil is not optimized for the 
high bands. 

##  If you compare tuner's  with the way PI or PI-L's  are designed and optimized  for
linear amplifiers, you will see a big difference !    Even Henry radio figured out this one. 
Henry radio would short out  1/2 the roller coil, with a cam arrangement, when the roller
coil  was used used on the upper HF bands.  A  small, separate 10m coil, usually wound using
flat strap, would be  mounted at right angles to the back end of the tapered pitch roller. 

##  IF  a bandswitched  arrangement is used,  it could be better optimized.   3/8"  or  1/2"
ribbon coils [or 1/4" to 3/8" tubing]   could be used on 160-30m....and more large diam
tubing could be used for  20-17-15m..and a flat strap coil for 10/12m....and all 3 x coils at
right angles to each other.    A 2-3 wafer  bandswitch, with all 3 x wafers strapped in parallel
could be used  for added current capability.   There is a trick to strapping parallel wafer's in 
any bandswitch.  If the coil taps are done via  wafer #1..... the output of the assy must be via
wafer #2. [ or #3  if  3 x wafer's used].     This  way, the  RF current will  divide equally 
between each wafer.  

##  In the above example,  IF the  coil  Q  could be increased to 250,  the power dissipated in the
coil would be 1/2.     If the coil[S] were optimized better, the UN loaded Q of the coil  could be 
a lot higher than 250, further reducing coil  diss.    In the Hb  T tuner's  I have seen that used
ceramic vac caps, the limitation was still the roller coil.  Even then, a 3/8"  ribbon coil  fares a lot
better  than using paltry 12 ga wire. 

##  It might be more productive to fix the ant in question,  right at the feedpoint, VS  building the ultimate tuner.  
Pretty simple matter  to make a 80m/or 160m dipole cover  both ssb/cw.   Rick's  idea of shunting a cap
at the feedpoint works well,  and will shift the resonance way up.   Other's have used relay switched coils
at the feedpoint instead, and that works  good too.[ shifts resonance down].     Some have used motor driven 
compressible coils,and that works very well, and no rolling/sliding contacts. 

Later.... Jim   VE7RF